Automatic target recognition methods of SAR images for military applications

Abstract

Automatic Target Recognition is an emerging developing field of newlinestudy that is crucial to military applications. Compared to optical images, the newlinemicrowave images captured by Synthetic aperture radar produce high newlineresolution images with a larger dynamic range. The automatic target newlinerecognition algorithm detects targets from images of Synthetic aperture radar newlinefollowed by a classifier to classify and recognize the target classes exactly. newlineThere are five stages in an Automatic Target Recognition process: newlinedespeckling, detection, classification, recognition, and identification. newlineHowever, an automatic target recognition system fulfills the first three phases newline(i.e) despeckling, detection and classification. This research work deals newlinesynthetic aperture radar image processing by preprocessing, detecting, newlineclassifying and recognizing the military vehicle targets from SAR images. newlineThe possible coherent picture acquisition system is synthetic newlineaperture radar. Because of multiplicative speckle noise, it is not easy to to newlineassess images captured by synthetic aperture radar. Consequently, newlinedespeckling images is the most important task in image processing for speckle newlinereduction. While speckle reduction is the main focus of all currently available newlinedespeckling techniques, not all will maintain the image edges. In the first newlinemodule of the research work, a despeckling algorithm using an Optimized newlineSparse Fuzzy Wavelet Transform with Minibatch Water Wave Swarm newlineOptimization is proposed for despeckling of SAR images. The experiment newlineoutcomes of the experiment demonstrate that the suggested transform gives newlinesuperior results compared with the existing methods by considering Root newlineMean Square Error(RMSE),Peak Signal-to-Noise Ratio, Structural Similarity newlineIndex Metrics(SSIM), and Pratt s FOM. newline